NOTHING needs to be explained…..everything clear as water.
Thank you so much guys, keep up the GREAT work

By default the OS will automatically fsck the filesystem “every 25 mounts or 180 days, whichever comes first”. This could be a serious issue if you want to reboot the system 180 days from now, as your 4 minute reboot could turn into a 4+ hour wait for fsck to complete.

If you want to disable the automatic fsck upon boot, try this command:

# tune2fs -c 0 /dev/sdb1
tune2fs 1.41.12 (17-May-2010)
Setting maximal mount count to -1

`tune2fs -i 0 /dev/sdb1` will also disable the automatic fsck.

-= Stefan

I’ve just found the cause of this problem.
http://www.linuxquestions.org/questions/showthread.php?p=4519659#post4519659
Maybe reply #9 explains the reason.
Could anyone tell me whether this issue has already resolved or not.

DMC

Thanks for the reply. While in Parted, I still can’t get anywhere because the disk’s label remains unrecognized, and the mklabel command simply returns the above-mentioned error (pretty much every command in Parted returns the same error: “unrecognised disk label”). Should I be using a different Linux tool to look at the drive?

I seem to have just purchased the same drive you are talking about, and yes I had the same problem. I ended it up there by making an error in following the instructions above. If you look at the drive, you will notice two partitions on it. Both zero in size. Use rm command for each partition and go carefully back to the beginning.
Note, though, that the above instructions result with some empty space at the beginning of the drive. If you select units as Bytes, the empty space will be moved to the end of the drive.

In my case the drive is a Hitachi XL3000 3TB (external USB 2.0) and I wanted to format it ext3. Since Parted didn’t work for me, I partitioned it under OS X (Disk Utility) as a single 3TB partition. And then mkfs.ext3 and tune2fs -L for the label. Seems to be working now, but if someone smarter than me wanted to explain how to do it without resorting to OS X, that would be awesome.

Any help?

I’ve updated the tutorial with step by step instructions.

If you do not add that to the command, you will loose (i think) either 5 or 10% of the drive to reserved blocks.

# mkfs.ext3 -m0 /dev/sda1

I applied mklabel gpt successfully, but I still see my RAID5 as 2T disk.

Although not well documented, parted will accept starting location and size in sectors, if values in MB are being rounded and causing slightly different values than desired. To repartition the fourth disk sdd (and lose all of its content):
parted -s /dev/sdd mkpart fat16 34s 1048609s

Note that I’m worried about duplicating the GUID that uniquely identifies each partition, not the standardized GUID value that identifies the partition type.

parted /dev/sda
mklabel gpt
mkpart primary 0 -0
print
quit
mkfs.ext3 /dev/sda1

SassyZ needs to understand that file systems are created within partitions while GPT are applied to disks. In other words, “parted mklabel gpt /dev/sda” writes a partition table to the first few sectors of a disk drive. The “mkpart” command creates a partition entry in that GPT. The first partition is named “/dev/sda1″. So the mistake in the last command is specifying the wrong device special file name. The command should be “mkfs.ext3 /dev/sda1″ to create a file system in the first partition. The command “mkfs.ext3″ will create a file system starting at sector zero and overwrite the GPT. (Note I’ve never tried a size of -0, so use print to verify the partition that was created.)

You really should run parted like so…

parted -a optimal /dev/sdb

The backup GPT table is not at the end of the disk, as it should be. This might mean that another operating system believes the disk is smaller. Fix, by moving the backup to the end (and removing the old backup)?

One smaller problem I’m facing now is also GPT-related:
I installed ubuntu on 3x2TB hw raid whilst there still was a 5X2TB soft raid in the same machine. Next I moved the data from the soft raid to the hw raid and expanded the hw raid to its full 8x2TB in raid 6.
When now I want to enlarge /dev/sda3 (ext4) to full raid size I get blocked of by the GPT backup table <>

But I don’t quite see how to do that.

Regards,
Jord

One question though… All this assumes you’ve got a running system to which you plan to add a second >2TB device. When using GPT, most modern linux distro’s will do as you described.

However, how do you install to such a device?

Situation = 8x2TB raid6 (hardware), and boot-able debian 6.0 cd -> The installer recognizes a 500 GB disk (I previously GPT’ed the device & partitioned it to my liking with gparted livecd).

Suggestions?

TIA,
Jord.

Number Start End Size File system Name Flags
1 17.4kB 3999GB 3999GB ext3 primary

any suggestions?